Comment by dekhn

It's pretty common to transfer copyright of the final manuscript to the publisher, while retaining a non-copyright pre-submission manuscript that is widely circulated. I don't know if this has ever been tested legally. I suspect Elsevier and others are trying not to litigate this heavily because they know the press and public will hammer them on it.

My postdoc advisor would receive the copyright transfer form from the publisher, modify the text to say he retained copyright, sign that, and send it back. Without fail, the publishers accepted that document, and published the paper. Again, I don't think this is legally tested, and my advisor said it's likely they didn't even notice the rewording of the copyright transfer document.

I thought the web would change this, but in my experience, people don't weight papers published in arxiv.org nearly as high as work published in peer-reviewed journals. And the vairous attempts at post-review (faculty of science, etc) haven't been able to replace the peer-reviewed journals successfully.